Re: Oracle Object Types and Java (migrating from mod_plsql)

In article <1108961046.813638.205280_at_c13g2000cwb.googlegroups.com>, Thanatos says...
>
>>bottom line, the htp stuff isn't hugely expensive here -- why not
>measure it?
>>See what amount of time you spend doing
>
>>a) sql
>>b) plsql
>
>Thanks for your ideas. But now I don't feel like getting up in the
>morning because of the asteroid :-)
>
>I guess I can prepare 2 versions of a page in our application.
>Typically a "results" page.
>
>v1: Generates the page with all necessary sql but not actually serve it
>via htp
>
>v2: The standard page which generates and serves it via htp.
>
>Assuming they are called:
>procedure results
>procedure results_nohtp
>
>Would you recommend as a test bed the approach you outline here:
>http://asktom.oracle.com/pls/ask/f?p=4950:8:9902943510168534793::NO::F4950_P8_DISPLAYID,F4950_P8_CRITERIA:668624442763
>
>eg:
>ops$tkyte_at_ORA9IR2> exec runStats_pkg.rs_start;
>ops$tkyte_at_ORA9IR2> exec p1
>ops$tkyte_at_ORA9IR2> exec runStats_pkg.rs_middle;
>ops$tkyte_at_ORA9IR2> exec p2
>ops$tkyte_at_ORA9IR2> exec runStats_pkg.rs_stop(1000);
>
>
>Regards
>

That would be a start, yes -- but you would be working with 'best case' for p1 since plsql would be doing all of the fetching -- rather than a back and forth over the network. but it would give you an idea.

(sql_trace would too -- you would see all of the sql's and their cpu times in the trace vs the time to generate the page)

-- 
Thomas Kyte
Oracle Public Sector
http://asktom.oracle.com/
opinions are my own and may not reflect those of Oracle Corporation